(dissenting).
I would affirm for substantially the reasons given by the trial judge in his oral opinion. Strickland v. Washington, 466 U.S. 668, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984) requires that two prongs be met to establish a claim of ineffective assistance of counsel. First, the defendant must show that counsel’s performance was deficient. Second, the defendant must show that the deficient performance prejudiced the defendant.
As to the first prong, the majority concedes only that counsel “should have interviewed Caldwell prior to having her testify.” But the trial judge found that “[t]his error was so serious as to fall in this instance below what is expected ... under the Sixth Amendment and the New Jersey Constitutional guarantee of effective assistance of counsel.” I agree with that assessment and would, therefore, expressly conclude that the first prong of Strickland was met.
It is a clear error of professional judgment in a criminal case not to interview an important alibi witness before that witness takes the stand.1 Counsel had two chances to interview Caldwell. *249The first opportunity came on Thursday, February 1, 1996, when she appeared at the courthouse but was not reached by the time of evening recess. The second chance occurred when she belatedly appeared after lunch on Tuesday, February 6, just before she was called to the stand.
It is true, as the State argues, that counsel is entitled to rely on the representations of his client in developing defenses and in deciding what witnesses to interview or not to interview. State v. Sloan, 226 N.J.Super. 605, 615, 545 A.2d 230, certif. denied, 113 N.J. 647, 552 A.2d 171 (1988). But such reliance does not extend to calling an alibi witness without interviewing her. So potent is alibi testimony that not to be certain as to precisely what such a witness will say is to invite the serious prejudice that occurred here.
Even assuming that the benefits of an alibi defense were worth the risk of mentioning it to the jury even though counsel was unable to discuss it with Caldwell prior to the start of trial, I also conclude that the judge was correct in finding it deficient performance not to have confirmed the details of the alibi with the witness when defense counsel had the opportunity to do so. Counsel could have conducted a brief interview when Caldwell appeared on Thursday, or he could have asked the trial judge for an opportunity to do so just before she took the stand the following Tuesday. I am not unmindful that the patience of the court and jury had been sorely strained at that point in the trial by Caldwell’s tardy appearance. But I nevertheless deem it a significant omission that counsel did not at least try to seek a brief recess to interview her.
Second, the failure of Caldwell to place the defendant with her at the time of the crime prejudicially damaged the claim of misidentification by Carter. Had the jurors harbored doubts *250about Carter’s testimony or questions concerning her identification of Banks, those doubts and questions surely paled in the face of the failed alibi.
The defense rested on the doubt raised by a very stressed victim’s identification, the failure to locate the gun and the jewelry and, most importantly, the alibi. When the alibi failed the whole defense case lost credibility. I respectfully disagree with the majority’s speculation that had counsel interviewed Caldwell, she probably would have been called as a witness anyway. Its view is that Caldwell’s testimony, notwithstanding her concession that she was not with Banks at 8:30 a.m., could still raise a doubt as to Banks’ participation in the crime, because Caldwell was not paying particular attention to the time, they had just arrived in Newark perhaps as late as 8:00 a.m., and then went to Banks’ house. But they then separated and Caldwell left for an appointment. She only surmised that Banks remained at his house and she specifically denied they were together at 8:30 a.m. If counsel had been able to interview her before the start of trial, his opening would not have so depended on alibi. If he had interviewed her before she testified, had she been called at all, her interrogation on the stand would have been quite different.
In contrast to the majority, however, I see little chance that had she been interviewed she would have been called. Without Caldwell’s testimony that she and Banks were together at 8:30 a.m., she adds little of probative value to the defense case. In fact, her admission that they had separated by 8:30 a.m. not only seriously detracts from viable defense arguments regarding the victim’s identification, but also places Banks near the scene of the robbery before it occurred, something none of the State’s other witnesses did.
Had the defense of alibi not been mentioned to the jury and Caldwell not been called to the stand, doubt could well have been engendered in the minds of some jurors as to the victim’s identification as the result of (1) the stress that Carter, the victim, was under; (2) the evidence of the suspect’s changes in clothing and *251hairstyle between the robbery and the chance sighting of Banks later in the morning; (3) the failure to find the gun and jewelry; and (4) the improbability of a robber returning to the scene so quickly after the crime — facts urged upon the jury during the opening. But when Caldwell’s testimony unexpectedly did not cover the time of the robbery and it happened that his house was within a few blocks of the crime, reasonable doubts about the State’s case in the minds of jurors could not have but been dispelled to Banks’ great prejudice. Thus, the second prong of the Strickland standard is established.
See Leonard N. Arnold, New Jersey Practice, § 961 (1999-2000) ("Interview all witnesses.”); John S. Applegate, Witness Preparation, 68 Tex. L.Rev. 277, 287 (1989) ("The practical literature uniformly views the failure to interview witnesses prior to testimony as a combination of strategic lunacy and gross negligence."); Steven Lubet, Modem Trial Advocacy: Analysis and Practice at 79 (2nd ed. 1997) ("[I]t is generally considered incompetent for a lawyer to fail to meet with and prepare a witness in advance of offering her testimony.”); U.S. v. Brant, 1993 WL 313369 (E.D.Pa.1993), aff'd, U.S. v. Snead, 27 F.3d 560 (3rd Cir.1994), cert. denied, Snead v. U.S., 513 U.S. 909, 115 S.Ct. 278, 130 L. Ed.2d 194 (1994). (finding that an attorney who cross-examined prosecution eyewitnesses without first having interviewed those witnesses had violated "an ancient principle of cross-examination, recognized by advocates for millennia: never ask a question to which you do not already know the answer.”) Most of the decided cases involving ineffective assistance claims that are premised on faulty witness preparation ultimately rest upon the failure of defense counsel to interview or call witnesses proposed by the defendant. Most of these claims fail because the defendant cannot establish prejudice in not calling any particular witness. See e.g. State v. Bey, 161 N.J. 233, 736 A.2d 469 (1999), cert. denied, Bey v. New Jersey, 530 U.S. 1245, 120 S.Ct. 2693, 147 L. Ed.2d 964, (2000). The case, sub *249judice, is gratifyingly rare because very infrequently does an uninterviewed friendly witness take the stand and proceed to so deeply wound the very side for which the witness was called.